Although evolutionarily just as ancient as IgM, it has been thought for many years that IgD is not present in birds. Based on the recently sequenced genomes of 48 bird species as well as high-throughput transcriptome sequencing of immune-related tissues, we demonstrate in this work that the ostrich (Struthio camelus) possesses a functional δ gene that encodes a membrane-bound IgD H chain with seven CH domains. Furthermore, δ sequences were clearly identified in many other bird species, demonstrating that the δ gene is widely distributed among birds and is only absent in certain bird species. We also show that the ostrich possesses two μ genes (μ1, μ2) and two υ genes (υ1, υ2), in addition to the δ and α genes. Phylogenetic analyses suggest that subclass diversification of both the μ and υ genes occurred during the early stages of bird evolution, after their divergence from nonavian reptiles. Although the positions of the two υ genes are unknown, physical mapping showed that the remaining genes are organized in the order μ1-δ-α-μ2, with the α gene being inverted relative to the others. Together with previous studies, our data suggest that birds and nonavian reptile species most likely shared a common ancestral IgH gene locus containing a δ gene and an inverted α gene. The δ gene was then evolutionarily lost in selected birds, whereas the α gene lost in selected nonavian reptiles. The data obtained in this study provide significant insights into the understanding of IgH gene evolution in tetrapods.
Immunoglobulins are defense molecules that are critical for animal adaptive immunity, and they are only present in jawed vertebrates, such as cartilaginous fish (Chondrichthyes), teleost (Osteichthyes), amphibians (Amphibia), reptiles (Reptilia, including birds and nonavian reptiles), and mammals (Mammalia) (1–3). These molecules are essential for driving B cell development and providing immunological protection in the context of humoral and mucosal immunity against invading pathogens. Different IgH (IgH chain) classes, also known as isotypes (determined by the H chain C region), perform largely distinct functions in pathogen neutralization, opsonization, and complement activation. Furthermore, IgH isotypes have also been selected by evolution to effectuate either systemic or mucosal immunity or mediate the allergic response. Comparative studies of vertebrates have identified >10 genes encoding different IgH classes. In addition to the widely known IgM, IgD, IgG, IgA, IgE, and IgY classes present in nonmammalian tetrapods (1–4), a number of species- or class-specific IgH classes have also been identified, such as IgO in platypus (5), IgX and IgF in amphibians (6–8), IgZ/T in teleosts (9, 10), and IgNAR in cartilaginous fish (11). In particular, the discovery of distinct IgH classes in jawed vertebrates has provided significant insights into IgH gene evolution, suggesting that IgM and IgD (also named IgW in lungfish and cartilaginous fish) are the ancestral IgH classes (8, 12).
IgM is the first IgH isotype to be expressed during B cell development and immunological responses. Membrane-bound IgM molecules form the BCR on the B cell surface, and BCR expression is essential for both B cell survival and the subsequent expression of other IgH classes. Targeted disruption of the μ gene in mice causes the arrest of B cell development at the pre-B cell maturation stage, thereby leading to B cell deficiency (13). Secreted IgMs, however, which include natural and immune IgMs, are involved in a variety of pathophysiologies, including infection, B cell homeostasis, inflammation, atherosclerosis, and autoimmunity (14). Most likely due to functional importance, IgM is highly structurally conserved and has been found in nearly all jawed vertebrates examined to date (15), the only exception being the African coelacanth, Latimeria chalumnae, which appears to possess only IgW (or IgD), but no IgM-encoding genes (16).
Although thought to be just as ancient as IgM in vertebrate evolution, IgD is a somewhat puzzling IgH isotype in terms of structure and function, as well as its distribution among jawed vertebrates (15, 17, 18). In humans and mice, expression of membrane-bound IgD is a developmental hallmark of mature B cells in peripheral lymphoid organs (19). Disruption of the δ gene in mice, however, causes only a mild phenotype with slightly delayed affinity maturation during the early primary response (20). Another study showed that in mice lacking an intact μ gene, forced IgD expression can largely substitute for the loss of IgM function in B cells (21). These data suggest that IgD is functionally redundant with IgM. Recently, it was shown that secreted IgD can bind to basophils through a calcium-mobilizing receptor that induces antimicrobial, opsonizing, inflammatory, and B cell–stimulating factors, thereby controlling an ancestral surveillance system that lies at the interface between immunity and inflammation (22).
In addition to these unique functions, IgD also differs greatly from IgM from an evolutionary viewpoint (15, 17). Since it was confirmed in 2006 that IgD is orthologous to IgW in lungfish and cartilaginous fish (12), this IgH isotype has now been identified in a large number of vertebrates. In contrast with IgM, which has a highly conserved four-CH-domain structure in jawed vertebrates, IgD exhibits a high degree of variation in the number of its CH domains, which ranges from 2 in mice up to 16 in zebrafish and 19 in sturgeon (9, 23, 24).
Of all jawed vertebrates that have been studied, only the African coelacanth lacks an IgM-encoding gene (16). As one of two most primordial IgH isotypes, IgD (or IgW) has been found in all examined fish, amphibians, and nonavian reptiles. However, analyses of IgH genomic loci clearly show that the IgD-encoding gene has been either “pseudogenized” or lost in several mammals, including rabbits, opossums, camels, elephants, and guinea pigs (25–29). More surprisingly, IgD has not yet been identified in any bird species, and there is clear evidence that the IgD-encoding gene is missing in chickens and ducks (30, 31). Therefore, it is commonly thought that the δ gene is absent in birds, contributing to the relatively small immune repertoire in this clade (32).
In addition to lacking the δ gene, in the IgH loci of both chickens and ducks, the IgA-encoding gene, α, is inverted and has an opposite transcriptional orientation to that of μ and υ (30, 31). Interestingly, a number of recent studies showed that the α gene is missing in certain nonavian reptiles, such as green anole lizards, snakes, and turtles, although these species all possess μ, δ, and υ genes (33–35). As nonavian reptiles are close relatives of birds, these data caused us to speculate that a genetic rearrangement event in the IgH locus may have occurred in the common direct ancestor of modern reptiles, which led to either loss of the δ gene and the inversion of the α gene in birds, or the loss of the α gene in nonavian reptiles (36). However, this hypothesis was strongly challenged by one of our recent studies on crocodilians, which showed that both δ and the inverted α genes are present in the IgH locus of this group (37). That study suggested that the loss and inversion of the α or δ genes were independent genetic events that occurred after the divergence of birds and nonavian reptiles, suggesting that the δ gene may be present in some birds.
The ostrich (Struthio camelus) belongs to the Ratitae order, which includes the most primitive living aves, having diverged from other avian lineages ∼140 million years ago (38). In a previous study, we identified only three IgH genes (μ, υ, and α) in the ostrich (36). However, based on high-throughput sequencing of transcriptome and IgH-specific transcripts derived from the bursa of Fabricius and the spleen, we demonstrate in this study that the ostrich has one δ gene and two additional μ and υ genes. Further analyses show that the common ancestors of modern birds should express multiple IgH isotypes, including IgD, as well as subclasses of IgM and IgY.
Materials and Methods
Animals, RNA and DNA isolation, and reverse transcription
Ostriches (S. camelus) were purchased from a local Beijing farm. The animals were handled in accordance with the guidelines of China Agricultural University regarding the protection of animals used for experimental and other scientific purposes. Total RNA from different tissues was extracted using the TRNzol kit (Tiangen Biotech). Reverse-transcription reactions were conducted using Moloney murine leukemia virus reverse transcriptase, according to the manufacturer’s instructions (Invitrogen). Genomic DNA was extracted from the liver using routine protocols.
Transcriptome analysis of the spleen and the bursa of Fabricius
Total RNA samples were derived from the spleen and the bursa of Fabricius and were mixed equally. The RNA library was examined using an Agilent 2100 Bioanalyzer (Agilent Technologies) and the ABI StepOnePlus Real-Time PCR System (Life Technologies), and then sequenced using an Illumina HiSeq 2000 platform (Illumina).
Total RNA was isolated from various tissues using an RNeasy Mini Kit (Qiagen), and RNA quantity and quality were assessed with a Nanodrop 2000 (Thermo Fisher Scientific). First-strand cDNA was synthesized using QuantiTect Reverse Transcription Kit (Qiagen) and oligo(dT)20 primers. Quantitative RT-PCRs were performed using LightCycler 480 SYBR Green I Master mix (Roche) under the following cycling conditions: 95°C 5 min, then 40 cycles of 95°C 10 s, 60°C 10 s, and 72°C 10 s. The primers used were as follows: EF1A1 (sense, 5′-TTT CTG GTT GGA ACG GAG AC-3′; antisense, 5′-ACG AGT TGG TGG CAG GAT AC-3′); IGHM1 (sense, 5′-TCA CCT TCT CCT GGA CCA AC-3′; antisense, 5′-CTG AGC CCA GCA GTA GAA GG-3′); IGHM2 (sense, 5′-TAC CTG CTC CGC AGC TAC TT-3′; antisense, 5′-GCT TTG GTG AGG AAG GTG TC-3′); IGHY1 (sense, 5′-CAC TGC CTC ATC TCC CAC TT-3′; antisense, 5′-CCT GGG CAC TTG TAG ACG TT-3′); IGHY2 (sense, 5′-ACC AAG TCC ATC GCC AAG-3′; antisense, 5′-CGT CAG CTT GCT GTA GAC GA-3′); IGHD (sense, 5′-GGC GGC AGC GTG GAG ATG TA-3′; antisense, 5′-GGC GAG CAG GTC GTC GTA GG-3′).
High-throughput sequencing analysis of IgH-specific 3′ RACE PCR products
RACE PCRs were conducted using JH-derived primers and total RNA derived from either the spleen or the bursa of Fabricius. PCR products were recovered and mixed for ion proton sequencing. An ion proton library was constructed using the Ion Fragment Library Kit (Life Technologies). The PCR products were fragmented using a Covaris LE220 machine (Covaris). The selection of DNA fragments (∼250 bp) was performed using a Tanon 2500 gel imaging system (Tanon) and purified using a QIAquick Gel Extraction Kit (Qiagen). The quality and concentration of the library were determined using an Agilent 2100 Bioanalyzer (Agilent Technologies) and the High Sensitivity DNA kit (Agilent Technologies). Emulsion PCR and all sequencing reactions were performed according to the ion proton protocol.
Southern and Northern blotting
The probes for Southern and Northern blotting were labeled with digoxigenin using the DIG-High Prime DNA Labeling and Detection Starter Kit (Roche). Hybridization and detection were performed according to the manufacturer’s instructions.
Long-range PCR amplification of genomic DNA fragments
The PCRs were performed using liver genomic DNA as template. The DNA polymerase used was LA-Taq DNA polymerase (Takara). All resulting PCR products were cloned into the pMD-19T vector (Takara) and sequenced.
DNA and protein sequence editing, alignments, and comparisons were performed using the DNASTAR Lasergene software suite. Multiple sequence alignments were performed using the Clustal W program (7984417).
The accession numbers of the sequences used in the construction of phylogenetic tree presented in this study were as follows: μ, Nurse shark IgM, I50731; Xenopus laevis IgM, AAH84123; Xenopus tropicalis IgM, AAH89670; Pleurodeles waltl IgM, CAE02685; lizard IgM, ABV66128; gecko IgM, ABY74509; Chinese soft-shell turtle IgM, ACU45376; red-eared slider IgM, AFR90255; Chinese alligator IgM1, AFZ39166; Chinese alligator IgM2, AFZ39167; Chinese alligator IgM3, AFZ39168; Siamese crocodile IgM1, AFZ39177; Siamese crocodile IgM2, AFZ39178; Siamese crocodile IgM3, AFZ39216; python IgM, AFR33764; ostrich IgM1, AFA41927; chicken IgM, P01875; duck IgM, CAC43061; gray goose IgM, AFM77858; pigeon IgM EMC81140; human IgM, AAS01769; mouse IgM, CAA24199; pig IgM, AAC48775; rabbit IgM, AAA64251; panda IgM, AAX73309; platypus IgM, AAO37747; α or χ, chicken IgA, AAB22614; duck IgA, AAK17834; human IgA, P01876; mouse IgA, AAB59662; pig IgA, I47175; cow IgA, AAC98391; sheep IgA, AAC64980; horse IgA, AAP80145; rabbit IgA, S09264; panda IgA, AAX73304; dog IgA, AAA56796; platypus IgA, AAL17700; X. tropicalis IgX, AAI57651; ostrich IgA, AFA41929; Axolotl IgX, CAO82107; δ, X. tropicalis IgD, ABC75541; lizard IgD, ABV75541; Chinese soft-shelled turtle IgD, ACU45375; red-eared slider IgD, AFR90257; Chinese alligator IgD, AFZ39188; Siamese crocodile IgD, AFZ39209; gecko IgD, ABY67439; human IgD, AAA52770; mouse IgD, P01881; panda IgD, AAX73311; platypus IgD, ACD31540; υ, X. laevis IgY, CAA33212; X. tropicalis IgY, AAH89679; Pleurodeles waltl IgY, CAE02686; Axolotl IgY, S31436; lizard IgY, ABV66132; frog IgY1, KN906073; frog IgY2, KN906073; snake IgY1, AFR33842; snake IgY2, AFR33843; Python IgY1, AFR33765; Python IgY2, AFR33766; Chinese alligator IgY1, AFZ39169; Chinese alligator IgY2, AFZ39170; Chinese alligator IgY3, AFZ39171; Siamese crocodile IgY1, AFZ39180; Siamese crocodile IgY2, AFZ39181; Siamese crocodile IgY3, AFZ39182; chicken IgY, CAA30161; duck IgY, CAD57004; gray goose IgY, AFN08753; ostrich IgY1, AFA41930. The nurse shark IgM gene was used as outgroups. The phylogenetic trees were constructed by the Bayesian Inference method (39) and viewed in FigTree version 1.4.1 (BEAST).
Analysis of ostrich IgH isotypes using high-throughput sequencing
Although three IgH genes (μ, υ, and α) were previously identified in the ostrich based on screening of an IgH-specific mini-cDNA library (36), it remained unclear whether the ostrich possesses a functional δ gene. To better explore bird macroevolution, we have recently sequenced the genomes of 48 modern avian species, including the ostrich (40). Using the crocodile δ gene as a template, we searched the ostrich genomic sequences for putative δ sequences and identified scaffold-807, which contains the ostrich α gene. In this scaffold, two CH-encoding exons that do not appear to belong to the known μ, υ, or α genes were identified (Supplemental Fig. 1). A BLAST analysis using the amino acid sequence deduced from these two exons against the National Center for Biotechnology Information database found that the highest sequence identities were to the crocodile δ gene. RT-PCR revealed that these two exons are transcribed in the bursa of Fabricius and the spleen, but not in any other tested tissues (data not shown).
These preliminary data suggested that there are more than three IgH isotypes expressed in the ostrich. To better characterize this group of genes, total RNA samples isolated from the bursa of Fabricius and the spleen were mixed together and subjected to transcriptome sequencing. A total of 49,193,508 clean reads with an average size of 100 bp was generated, accounting for ∼38,583 annotated unigene sequences. In addition to the previously identified μ, υ, and α genes, BLAST searches against these unigenes uncovered three more IgH genes, including IgD-, IgM-, and IgY-encoding genes.
As the transcriptome analysis was principally based on sequencing of total RNA isolated from a tissue with many cell types, functional genes might still be missed if they are only transcribed at very low levels. To ensure that all functional ostrich IgH genes are identified, we used ion-proton technology to further sequence the mixed 3′ RACE PCR products amplified with JH primers and RNA samples isolated from the bursa of Fabricius, spleen, and small intestine, the tissues in which IgH genes are thought to be most highly expressed. Theoretically, the majority of, if not all, 3′ RACE PCR products amplified using JH primers should represent IgH transcripts. Using this approach, ∼177 Mb of sequences were generated, which were assembled into six unique IgH gene sequences: two IgM, one IgD, one IgA, and two IgY H chain-encoding genes. These findings were consistent with the results from the transcriptome analyses, thus confirming that six functional IgH genes are expressed in the ostrich. We termed the newly identified μ and υ genes as μ2 and υ2, respectively, whereas the previously discovered genes were named μ1 and υ1. The identities of these genes were verified by phylogenetic analysis (Fig. 1).
Expression of μ2, a second functional IgM-encoding gene in the ostrich
As stated above, in addition to the previously identified gene μ1, a second functional IgM-encoding gene, μ2, is also expressed in the ostrich. Analysis of 3′ RACE PCR products suggested that the μ2 gene is primarily expressed in a secreted form. Interestingly, a comparison of the μ1 and μ2 genes revealed that they shared exactly the same CH4 sequence (Fig. 2), suggesting that a gene conversion or exon duplication event may have occurred between these two genes. The entire C regions of these two μ genes share an overall amino acid identity of 59.7%, whereas this value drops to 43.1% if CH4 domains are not included. No significant sequence differences were observed between these two genes in terms of cysteine distribution or putative N-linked glycosylation sites (Fig. 2). Quantitative PCR (Q-PCR) showed that in a 1-mo-old ostrich (Fig. 3A), both μ1 and μ2 genes were expressed relatively at higher levels in immune-related organs such as the spleen, the bursa of Fabricius, and the intestine, and the μ2 was expressed at much lower levels than the μ1 in these tissues. For the 1-y-old ostrich (Fig. 3B), the μ1 was still highly expressed in the above tissues, whereas μ2 was mainly expressed in the large intestine.
IgM subclass diversification is not common in tetrapods. However, it was recently reported that there are three functional μ genes in the IgH locus of crocodilians (37). As crocodilians are phylogenetically close to the ostrich, we performed a separate phylogenetic analysis using only the μ sequences derived from various tetrapods (Figs. 1, 4A). The results suggested that the divergence of the two ostrich μ genes occurred after the divergence of the ostrich and crocodilians.
Expression of υ2, a second functional IgY-encoding gene in the ostrich
Using the approaches described above, we also identified a second IgY-encoding gene in the ostrich, termed υ2. The υ2 gene also encodes a four-CH IgY H chain, which shows 49.8% sequence identity with υ1 at the amino acid level. Compared with υ1, υ2 shows a more conserved cysteine distribution relative to IgY-encoding genes in other birds and reptiles. For example, υ2, but not in υ1, features a conserved double cysteine in the N-terminal of CH1 and another cysteine in the N-terminal of CH2 (Fig. 5). Q-PCR showed that in the 1-mo-old ostrich (Fig. 3C), the υ1 gene was highly expressed in immune-related tissues such as the spleen, small intestine, and large intestine, whereas the υ2 was expressed at much lower levels than υ1 in all tissues. In the 1-y-old ostrich (Fig. 3D), υ1 was expressed at relatively higher levels in the spleen, large intestine, and thymus, whereas υ2 expressed at relatively higher levels in the spleen than in other tissues.
Expression of a functional IgD-encoding gene in the ostrich
In addition to the discovery of two μ genes, two υ genes, and an α gene, the most important finding in this study was the identification of a functional IgD-encoding gene in the ostrich. Based on δ sequences derived from our genomic survey and high-throughput sequencing, we performed 3′ RACE amplifications on the 3′ portion of transcribed ostrich IgD H chains. These efforts led to the cloning of a membrane-bound IgD H chain transcript. The full-length IgD H chain C region cDNA was amplified with primers derived from JH and the 3′ untranslated region sequences. The amplified full-length IgD H chain C region encoded 777 aa, consisting of seven CH domains and a transmembrane tail (Supplemental Fig. 2).
To confirm the existence of a δ gene in the ostrich genome, we performed Southern blots using ostrich genomic DNA with the IgD CH2-encoding exon as a probe (Fig. 6A). Our results showed that the δ gene was present in the ostrich genome, most likely as a single-copy gene. Considering that the 3′ RACE PCR amplifications identified only a single transcript encoding a membrane-bound form and that primers derived from the JH and 3′ untranslated region sequences amplified only a single PCR product from spleen or bursa of Fabricius total RNA samples, it is very likely that the ostrich δ gene encodes only a single membrane-bound product. To confirm this, we performed Northern blotting analyses using spleen mRNA samples to detect the expression of both the δ and μ1 genes (Fig. 6B). Only a single band of ∼3 Kb was detected for the δ gene, whereas two bands, corresponding to transcripts of the membrane-bound and secreted forms, were detected for the μ1 gene. Q-PCR results showed that the δ gene was expressed at much lower levels than the μ and the υ genes, although it was expressed at relatively higher levels in the spleen and the bursa of Fabricius compared with other tissues (Fig. 7).
The δ gene is located downstream of the μ1 gene and upstream of the inverted α gene in the ostrich IgH gene locus
To analyze the genomic organization of the ostrich IgH gene locus, we built an ostrich genomic bacterial artificial chromosome library consisting of 145,920 clones with an average insert size of 146 Kb. Therefore, this library has a ∼9.9-fold coverage of the ostrich genome. Unfortunately, we failed to obtain IgH gene-containing clones from this genomic library. Next, we employed a long-range PCR approach combined with sequencing to determine the genomic arrangement of the ostrich IgH genes. Results showed that the δ gene is located ∼1 Kb downstream of the μ1 gene, and the inverted α gene is located ∼5 Kb downstream of the δ gene (Fig. 8). The μ2 gene was shown to be located in a head-to-head orientation downstream of the α gene. We failed to map the locations of the υ1 and υ2 genes to this locus using long-range PCR, perhaps because they are located very far downstream of the μ2 gene.
The δ gene is widely distributed among bird species
Evidence supporting the lack of δ genes in birds was based on studies performed on just chickens and ducks (30, 31). However, the discovery of a functional δ gene in the ostrich indicated that δ genes might also be present in other bird species. Using the ostrich δ sequence as a template, we conducted a thorough genomic survey using BLAST searches against the 48 recently sequenced bird genomes. Strikingly, δ sequences were clearly identified in 18 bird species from 12 orders (Fig. 9, Supplemental Fig. 3), demonstrating the wide distribution of δ genes among birds and suggesting that the δ gene might only have been evolutionarily lost in selected bird species.
Divergence of μ1 and μ2, υ1, and υ2 occurred in common ancestors of modern birds
As the amino acid identities between μ1 and μ2 (with the CH4 domain not included for comparison), υ1, and υ2 are <50%, we suspected that the divergence of these Ig subclass-encoding genes may have occurred in the early stage of bird evolution. To address this question, we analyzed the IgH isotypes expressed in the Emu (Dromaius novaehollandiae), another primitive bird species, which is thought to have diverged from the ostrich ∼50 million years ago (41). Similarly, we used ion-proton technology to sequence the mixed 3′ RACE PCR products amplified with JH primers and RNA samples isolated from the bursa of Fabricius, spleen, and small intestine of the Emu, which showed that the emu also expressed six IgH isotypes, including two μ, two υ, one δ, and one α genes. Phylogenetic analyses revealed that both subclasses of the emu μ and υ genes were orthologous to those of ostrich, respectively (Figs. 1, 4A, 4B), and that the divergence of these subclass-encoding genes should have occurred before the divergence of the modern birds. As shown in the phylogenetic trees, the μ2 was obviously lost in most birds, although both the μ1 and μ2 genes are expressed by the ostrich and emu (Figs. 1, 4A). Although the ostrich and emu have two υ genes, many other birds only evolutionarily retain one of them (Figs. 1, 4B). These data strongly suggest that the common ancestors of modern birds should have at least six IgH isotype-encoding genes, including μ1, μ2, δ, υ1, υ2, and α, and some of these genes have been evolutionarily lost in modern birds (Fig. 10).
IgM and IgD are the first Igs expressed during B cell development, and the coexpression of these two Igs on the surface of B cells is regarded as a marker of mature B cells in mammals. Additionally, extensive comparative studies suggest that these two Igs are the most ancient IgH classes, emerging alongside the development of adaptive immunity in jawed vertebrates. However, for decades, it has been commonly believed that IgD is not present in birds, despite its wide distribution in mammals, nonavian reptiles, amphibians, and fish. In this study, we show that an IgD-encoding gene is present and functionally transcribed in the ostrich. Furthermore, a genomic survey of bird genomes revealed that homologous genes are also present in many other bird species. We also showed that, as in crocodilians, subclass diversification of IgM and IgY occurred in the ostrich.
Considering the identification of IgD-encoding genes in birds, IgD has now been found in all classes of jawed vertebrates. Although it is clearly absent in a few selected mammalian and bird species, this IgH class has been identified in all nonavian reptiles, amphibians, and bony fish species examined to date, suggesting, from an evolutionary point of view, that IgD is important in B cell development and/or the humoral response. Consistent with this idea is the fact that the African coelacanth appears to only express IgD(W) (16). However, despite this deep evolutionary conservation, IgD, compared with IgM, does show a great deal of variation with respect to its structure and expression patterns among different jawed vertebrates. Whereas IgD H chains possess no more than three CH domains in placental mammals, the number of CH domains varies from 2 to 11 in the expressed IgD H chains of the platypus and other tetrapods, including birds, modern reptiles, and amphibians (5, 8, 12, 33–35, 37). This variation is even greater in fish, with the rainbow trout having the shortest IgD H chains, with four CH domains (22205025) (42), and the Siberian sturgeon having the longest, with 19 CH domains (24). Variations in the number of CH domains in expressed IgD H chains can be attributed not only to varying numbers of germline CH-encoding exons, but also to extensive alternative RNA splicing events. In the Siberian sturgeon, nine IgD H chain transcripts with different numbers of CH domains were identified, and these different transcripts are apparently expressed from a same germline δ gene through alternative RNA splicing (24). Interestingly, alternative RNA splicing can also generate chimeric IgD H chains by splicing the μ CH1 domain onto IgD sequences, which has been observed in bony fish and pigs (43–45).
In placental mammals, IgD is primarily expressed in the form of membrane-bound B cell receptors, but it can also be expressed in a secreted form, albeit at very low levels (17). In this study, ostrich IgD was only detected in the membrane-bound form at the RNA level. Indeed, it appears common in tetrapods that IgD is primarily expressed in a membrane-bound form. For example, in platypus and lizards, only membrane-bound IgD-encoding transcripts can be detected (5, 33). In other species, such as snakes, turtles, crocodiles, and Xenopus, the IgD expression pattern is more like that in placental mammals (8, 12, 34, 35, 37). In all fish examined to date, membrane-bound IgD-encoding transcripts have been detected, and secreted IgD has been clearly detected at the protein level in selected species, including catfish and rainbow trout (46, 47). The concentration of IgD in catfish serum was estimated at 0.03 mg/ml (46), which is approximately equal to the concentration of IgD in human serum (48). Interestingly, in lungfish, the majority of IgD(W)- or IgD(W)-related IgH transcripts encoded secreted Igs (49, 50).
Despite variations in structure and expression, the discovery of δ genes in birds finally allows us to draw a clear picture describing the evolution of IgH genes in tetrapods. Based on all comparative studies of IgH genes in tetrapods, it is reasonable to speculate that the IgH gene locus in the common ancestor of tetrapods contained the μ, δ, α(χ), and υ genes, which were organized in the order μ-δ-α(χ)-υ (Fig. 10). In amphibians, such as X. tropicalis, the IgH gene locus essentially maintains this ancestral organization, with the addition of an IgF-encoding gene originally duplicated from the υ gene (8). In the common direct ancestor of birds and modern reptiles, the basic genomic organization of μ-δ-α(χ)-υ was maintained, although the α gene was inverted due to an unknown genetic rearrangement; this backbone structure can be found in both ostriches and crocodilians (37). Based on this genomic backbone structure, the α gene was then lost in specific reptiles, such as lizards (33), whereas the δ gene was lost in specific birds, such as chickens and ducks (30, 31) (Fig. 10). Generally speaking, all IgH genes in the common ancestor of tetrapods were inherited by mammals, but the υ gene became the IgG- and IgE-encoding genes through gene duplication and neofunctionalization (Fig. 10). Supporting this hypothesis is evidence derived from the most primitive mammalian species, the platypus, in which an υ-like gene, termed ο (IgO-encoding gene), was identified, in addition to the IgG- and IgE-encoding genes (5).
Another important finding of this study was the subclass diversification of μ and υ genes in the ostrich. Phylogenetic analyses suggested that the subclass gene diversification of both μ and υ genes most likely occurred during the early stages of bird evolution but after their divergence from nonavian reptiles. Interestingly, similar to the situation in crocodilians, the ostrich μ2 gene is located downstream of the α gene, whereas μ1 is positioned in the most 5′ region of the IgH C region gene locus. Therefore, the μ2 gene is most likely expressed through class-switch recombination, as we demonstrated in crocodilians (37). These class-switched IgM molecules may play significant roles in humoral immunity in ostriches and crocodilians.
In conclusion, we demonstrated that an IgD-encoding gene is functionally transcribed in the ostrich and is also present in many other bird species, thus invalidating the commonly held notion that δ genes are absent in birds. Additionally, this study also suggests that the common ancestors of modern birds should have at least six IgH isotype-encoding genes, including μ1, μ2, δ, υ1, υ2, and α, and some of these genes have been evolutionarily lost in modern birds. Our findings provide significant new clues to enhance our understanding of IgH gene evolution in tetrapods.
This work was supported by the National Natural Science Foundation of China (Grants 31272433 and 31472085).
The sequences presented in this article have been submitted to the National Center for Biotechnology Information's GenBank (http://www.ncbi.nlm.nih.gov/) under accession numbers KM510515, KM510516, KR030054, KU641022, KU641023, KU641024, KU641025, KU641026, and KU641027.
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.